Harlan's Impossible turbo build.
12-01-2012, 11:16 PM
#51
Having the best atomization is nice, but this is about flow control. Even using PWM controlling pumps the flow is not proportional. In fact with the sureflow pump everybody uses and a small nozzle the pump will cycle on and off at shutoff pressure even if the controller is only demanding 25%. I ended up running a recirc when I had that setup and that was the only way to get a measure of control. I want to be able to control flow from a trickle to a full spray and everywhere in between. This should allow me to have low flow while spooling and full flow when at full boost/power.
Update: Didn't get everything I wanted done today, but I managed to get the knock detection software working with my tone generator. But working on it gave me a migraine. It will detect 3.5 khz without a problem and give me a proportional output. Still have to decide if I want to feed it into the stock ecu or use my own knock retard scheme. Wish Oltmann was around for advice.
Update: Didn't get everything I wanted done today, but I managed to get the knock detection software working with my tone generator. But working on it gave me a migraine. It will detect 3.5 khz without a problem and give me a proportional output. Still have to decide if I want to feed it into the stock ecu or use my own knock retard scheme. Wish Oltmann was around for advice.
Last edited by Harlan; 12-01-2012 at 11:48 PM.
12-02-2012, 02:23 PM
#54
The current plan is to use ion sense to detect knock. It will give a 0-5 volt signal based off resistance of the spark gap. It's almost identical to the signal that the regular knock sensor sends but it is only influenced by cylinder pressure not other mechanical noise. Take a look at this:
http://www.google.com/url?sa=t&rct=j...2vc0En9QDMVJwQ
It's really pretty simple, uses a zener diode a capacitor a regular diode and a couple resistors. Haven't tried it out yet because it relies upon a coil with a separate primary and secondary ground, and that was why I started my ignition mod.
Since it should detect VERY low levels of knock the actual retard signal will vary based on amount of knock. In this situation it might be beneficial to let the ignition ride in a low level knock state and then just maintain it.
Edit: Also it will allow for full ignition cut when knock reaches a pre-determined level. This should be the ultimate failsafe and prevent significant engine damage if all else fails.
This isn't theoretical, it's in practice, but it's my first time trying it. Right now I can either feed an audio signal into the ecu knock sensor lines, or I can modify the ignition timing myself between the ecu and the coils themselves. I'd rather let the ecu handle it, but I don't know how good it would be or even if it's set up right for our cars.
I think the problem with the stock ecu is that it filters (analog or digital) the frequency our engines make (around 3.5khz) and only listens for normal piston knock (6-7khz). But I haven't gotten off my butt and fed the ecu tones to find out. The stock knock sensor is pretty much a microphone it doesn't care what the frequency is so the problem is somewhere else.
Oh and for WI percentage, I want to ramp from 25% to 50% water to fuel. Beyond that water becomes much less effective then water/meth.
http://www.google.com/url?sa=t&rct=j...2vc0En9QDMVJwQ
It's really pretty simple, uses a zener diode a capacitor a regular diode and a couple resistors. Haven't tried it out yet because it relies upon a coil with a separate primary and secondary ground, and that was why I started my ignition mod.
Since it should detect VERY low levels of knock the actual retard signal will vary based on amount of knock. In this situation it might be beneficial to let the ignition ride in a low level knock state and then just maintain it.
Edit: Also it will allow for full ignition cut when knock reaches a pre-determined level. This should be the ultimate failsafe and prevent significant engine damage if all else fails.
This isn't theoretical, it's in practice, but it's my first time trying it. Right now I can either feed an audio signal into the ecu knock sensor lines, or I can modify the ignition timing myself between the ecu and the coils themselves. I'd rather let the ecu handle it, but I don't know how good it would be or even if it's set up right for our cars.
I think the problem with the stock ecu is that it filters (analog or digital) the frequency our engines make (around 3.5khz) and only listens for normal piston knock (6-7khz). But I haven't gotten off my butt and fed the ecu tones to find out. The stock knock sensor is pretty much a microphone it doesn't care what the frequency is so the problem is somewhere else.
Oh and for WI percentage, I want to ramp from 25% to 50% water to fuel. Beyond that water becomes much less effective then water/meth.
Last edited by Harlan; 12-02-2012 at 07:40 PM.
12-04-2012, 05:14 AM
#58
I divide by zero
Join Date: Aug 2007
Location: Spring Hill, FL
Posts: 1,192
Likes: 0
Received 0 Likes
on
0 Posts
12-04-2012, 07:45 AM
#59
i have been reading some nasa and military papers concerning water injection.From what I have been reading and from what the OP has been posting I think he is going for a lean burn with cooler exhaust temps.
At 50% water many other factors start coming into play than what the usual w/m system discussions cover. At least from what I can understand.
At 50% water many other factors start coming into play than what the usual w/m system discussions cover. At least from what I can understand.
12-05-2012, 08:04 PM
#61
Since there's a lot of confusion here. Let me be plain about my current(tm) plans.
A lot of info for my build has come from Rotorocks build. Without it I would not have decided to run a 2" charge pipe.
https://www.rx8club.com/major-horsep...r-turbo-97855/
I'm running a 2" charge pipe to the front. Because of the length and small size it will in part act as an intercooler. It will be tucked up by the drive shaft so ground clearance will not change ( some country roads suck.)
The reason I'm using modified glasspacks is because I want to keep the thing street legal, and I don't want it to **** off the neighbors too much. If it's still too loud I'm gonna add removable baffles.
The exhaust is going to be wastegate prefered. If I sized it right the turbo should spool fast then the WG takes over. Spring is set at 4psi. I will probably have to go for a tighter spring even with boost controler, but it's a safe place to start.
I want to have my ignition mod finished because I want a working knock detection. It's progressing very nicely, but if it falls through I will use the stock knock sensor for both a readout/headphones and to feed the ECU a signal that will trigger it.
IAT will be taken from the same spot before the TB so the ECU will have some chance to save the engine if things go south. Also it will be big for safety while tuning.
Water injection is well proven. Preturbo water injection is also well proven. Although people do claim turbine damage the people who run it don't care because the benefits are huge. I don't think I'll have regrets. 50% water to fuel is a LOT, but I may need that much to keep my temps down/prevent knock. I'd rather be oversized for my needs and change software than be undersized and have to redo.
I know everyone is looking for the same things out of a turbo build, and has an unlimited budget with no other concerns. I also understand how all the turbo kits for our car are high quality and work properly without modification. But I'm gonna go out on a limb here and do my own thing and do it while I have the time and the money available to do it. Major modification begins in 8 days. I'll update on anything big until then.
A lot of info for my build has come from Rotorocks build. Without it I would not have decided to run a 2" charge pipe.
https://www.rx8club.com/major-horsep...r-turbo-97855/
I'm running a 2" charge pipe to the front. Because of the length and small size it will in part act as an intercooler. It will be tucked up by the drive shaft so ground clearance will not change ( some country roads suck.)
The reason I'm using modified glasspacks is because I want to keep the thing street legal, and I don't want it to **** off the neighbors too much. If it's still too loud I'm gonna add removable baffles.
The exhaust is going to be wastegate prefered. If I sized it right the turbo should spool fast then the WG takes over. Spring is set at 4psi. I will probably have to go for a tighter spring even with boost controler, but it's a safe place to start.
I want to have my ignition mod finished because I want a working knock detection. It's progressing very nicely, but if it falls through I will use the stock knock sensor for both a readout/headphones and to feed the ECU a signal that will trigger it.
IAT will be taken from the same spot before the TB so the ECU will have some chance to save the engine if things go south. Also it will be big for safety while tuning.
Water injection is well proven. Preturbo water injection is also well proven. Although people do claim turbine damage the people who run it don't care because the benefits are huge. I don't think I'll have regrets. 50% water to fuel is a LOT, but I may need that much to keep my temps down/prevent knock. I'd rather be oversized for my needs and change software than be undersized and have to redo.
I know everyone is looking for the same things out of a turbo build, and has an unlimited budget with no other concerns. I also understand how all the turbo kits for our car are high quality and work properly without modification. But I'm gonna go out on a limb here and do my own thing and do it while I have the time and the money available to do it. Major modification begins in 8 days. I'll update on anything big until then.
12-05-2012, 08:55 PM
#62
I am sure interested in that knock detector system you are putting together.
Are going for one on each rotor like the S2 models?
With such low boost I dont think you will need 50% water for detonation protection--unless you are going for a very lean burn.
Are you concerned about droplet separation with that high of percentage and the distance involved?
Here is a cool pic--put it in your trunk!:0
http://i147.photobucket.com/albums/r...d/IMG_0023.jpg
Are going for one on each rotor like the S2 models?
With such low boost I dont think you will need 50% water for detonation protection--unless you are going for a very lean burn.
Are you concerned about droplet separation with that high of percentage and the distance involved?
Here is a cool pic--put it in your trunk!:0
http://i147.photobucket.com/albums/r...d/IMG_0023.jpg
12-05-2012, 10:05 PM
#63
Yeah, if I can get knock detection working it will be the holy grail of rx8 tuning. I'm really amazed nobody tried the method I found to put det phones on the stock knock sensor. What I have planned will give individual knock indication for each rotor, and give individual retard to each rotor. We'll see if it works... 4psi is without a boost controller. I do plan on running higher, but I should be able to work the kinks out of it at low pressures. Yes 50% is insane for that low a boost... But at 10psi it might get more reasonable when running high compression with no intercooler. Droplet separation shouldn't be an issue until it hits the UIM because of the velocity. After that, who knows. But the cooling has already happened for the charge, and as long as it separates between the rotors evenly it shouldn't matter. If it doesn't I'll see the difference in knock. (assuming that works)
Lean burn is something I've been waiting for a turbo to pursue again. Not a central issue for this build, but something I'm leaving room for.
Edit: Studied up on Subaru knock detection. The system looks good and better designed than anything I could cook up. I does not differentiate between knock volume, although it does have different threshold at different rpms. I'm gonna feed it until it retards timing, then use that volume and frequency to tie in my own system. If someone is better at EE than me it should be pretty easy to attach a filter and frequency multiplier to the stock sensor to allow it to work properly, but that's beyond my goals.
Lean burn is something I've been waiting for a turbo to pursue again. Not a central issue for this build, but something I'm leaving room for.
Edit: Studied up on Subaru knock detection. The system looks good and better designed than anything I could cook up. I does not differentiate between knock volume, although it does have different threshold at different rpms. I'm gonna feed it until it retards timing, then use that volume and frequency to tie in my own system. If someone is better at EE than me it should be pretty easy to attach a filter and frequency multiplier to the stock sensor to allow it to work properly, but that's beyond my goals.
Last edited by Harlan; 12-06-2012 at 07:16 PM.
12-28-2012, 05:30 PM
#64
Well like every previous turbo build It's been almost a month without update, and my original time table has gone out the door. But I got a lot done today, and if I want to drive again on Monday, I better get more done tomorrow. Here's a look. Still deciding how I'm gonna mount everything. Hopefully by Saturday I'll have everything back in place with the turbo blank flanged and the wastegate spring pulled.
12-28-2012, 08:32 PM
#66
rev it up
I hope you not using the PVC bend in your intake as it will distort in the heat and cause problems. I used a screen that was mounted in a 3"PVC pipe and it was difficult to clamp and caused leaks.
I have removed this component and the car runs a lot better now.
I have removed this component and the car runs a lot better now.
01-13-2013, 05:04 PM
#67
PVC start to become pliable somewhere around 150 and can be bent to shape around 200-250. The back of the car should be cool enough to keep the pvc from melting as long as I keep the radiant heat off. PVC was not my first choice, but it's cheap and easy to fabricate. Gosh family and flu slow things down badly, but I have had time to get the car roadworthy again. Here's what I have so far.
01-14-2013, 06:12 PM
#69
Got the exhaust on and the car started. Not too terribly loud, but a bit raspy. I pretty much expected this from listening to all the rx7s with glass packs, hopefully when the turbo is aligned it will take a lot of the remaining noise and rasp out of the exhaust. Gonna get some help tomorrow patching any holes in the exhaust and welding the charge tube. Still need to work on oiling before I can even align the turbo and go start using the new intake/calibrating the new MAF. At this point I'll be happy to just be painting the exhaust header and wrapping it.
The coding an arduino as a boost controller is pretty simple. I have a GM 3bar map which is pretty much linear from 30" vac at 0v to 2bar at 5vols. It takes a 12volt input and the 0-5volt output can be sent straight to the arduino analog in. For boost control I'm using a 12volt 3port MAC solenoid. At 30hz pwm the arduino can control it pretty well. Right not I have it coded so that at a certain psi (voltage) it starts opening the wastegate and then at a higher psi it places the wastegate full open. I plan on having it read a rpm based table and control at different pressures at different RPMs. When I have the code a little more polished I'll be more than happy to share it. Right now it's pretty rough because I don't have it on the car. I'll probably throw in a high pressure ignition cut eventually.
The coding an arduino as a boost controller is pretty simple. I have a GM 3bar map which is pretty much linear from 30" vac at 0v to 2bar at 5vols. It takes a 12volt input and the 0-5volt output can be sent straight to the arduino analog in. For boost control I'm using a 12volt 3port MAC solenoid. At 30hz pwm the arduino can control it pretty well. Right not I have it coded so that at a certain psi (voltage) it starts opening the wastegate and then at a higher psi it places the wastegate full open. I plan on having it read a rpm based table and control at different pressures at different RPMs. When I have the code a little more polished I'll be more than happy to share it. Right now it's pretty rough because I don't have it on the car. I'll probably throw in a high pressure ignition cut eventually.
01-14-2013, 09:10 PM
#70
getting the grunt work out of the way...........
? what happens to the water/meth if the wastegate is open? You are still planning on injecting pre turbo--right?
i ran across this--I thought it may be of some interest:
http://www.google.com/url?sa=t&rct=j...G7sPmPmf5T13pQ
? what happens to the water/meth if the wastegate is open? You are still planning on injecting pre turbo--right?
i ran across this--I thought it may be of some interest:
http://www.google.com/url?sa=t&rct=j...G7sPmPmf5T13pQ
01-15-2013, 10:49 AM
#72
Wastegate open just vents the exhaust off the drivers side tailpipe. I think you meant BOV. I'm planning on recircing it so it shouldn't be a problem, hopefully it will be far enough downstream of the MAF to avoid fouling it. I love the mechanical water injector. The air atomizing valve using boost pressure is an elegant solution. However it runs into the same problem as everything else, in order to get flow control you need a pwm valve feeding the water to it. If you want on/off control it's the best you can get.
01-15-2013, 04:05 PM
#73
j_tibo
Join Date: Feb 2011
Location: Ottawa...well Wendover
Posts: 78
Likes: 0
Received 0 Likes
on
0 Posts
Nice to see your progress but I would take a second look at those welds that wont hold up very long before it leaks bad. Best thing to do would be to tack together then get a shop or a friend with a tig to make a much better and stronger well. Almost looks like you welded with fluxcore.
Cheers, good luck!
Cheers, good luck!
01-15-2013, 05:20 PM
#74
There are just so many thing I am shaking my head at I don't know where to start there are reasons I chose the parts I did and why I shake my head at some of the things I did in my original build thread. I think its cool to try something new, but being inventive only goes so far without looking at the basics.
